Fermi is the name of the architecture for NVIDIA's next gen (DX11) cards. Fermi was announced ahead of actual card announcements or even just information about gaming features. All that was talked about, in fact, was Tesla-related shit, but despite that I've read all kinds of bullcrap from people jumping to all kinds of ridiculous conclusions about it.
Once again, this was an announcement for Tesla. Companies looking to make large investments in new servers and HPC systems need a lot of lead time to make decisions, and NVIDIA was trying to appeal to them, as well as investors and stock holders, proving that Fermi is real and that there are some really cool things to look forward to about it. AMD released their shit, so now NVIDIA wants to make some sort of response, even if it isn't actual hardware. This was an announcement to gain mindshare, nothing more.
You know, with the impending stagnation of the graphics market at hand, I kind of thought the whole fanboy wars thing would more or less die off, as it really didn't matter anymore who you chose. But no, I was misguided. It seems that as performance has become adequate for almost all games now and in the foreseeable future, the graphics card market has become much more an exercise in the politics of sports than it used to be, with fans rooting for their "home teams", more out of senseless preference than anything else, because, after all, you can buy any card and still play any game you want.
Now I'm not without my biases. As I've stated here in the past, I have an affinity for NVIDIA. PhysX may not be God's gift to gamers, but it's the only GPU physics being used in games, and that's becoming important. Also, there are some driver features I'd be reluctant to leave behind. Like, for instance, being able to have automatic game profiles without friggin having to download ATITool. Come on ATI! No one's going to assume you're cheating because you make your drivers more intelligent and convenient. They already assume you're cheating, at least to some extent, however minor, just like it's assumed of NVIDIA and everyone else! What's the rationale here? This isn't 2003 anymore.
Plus the fact that CUDA has actually taken off to a much greater degree than any other equivalent is a plus for NVIDIA. Sure, one has to admire AMD's push for open standards, but unfortunately this is still 2009, not 2011, and we have to deal with the realities that affect us today. NVIDIA has paid more people, pushed more people, and just gotten more people to use their standard, and with future software tools like Nexus (or Nexem? or whatever), this will only continue to expand. It's not as though such a thing is unheard of, just look at x86. That's really what Fermi is all about. Seizing whatever opportunities NVIDIA has or has created for themselves in the general purpose world. That was their emphasis during their launch this week and should be the focus of enthusiasts, but unfortunately, that's not how things have turned out.
People are jumping to all kinds of conclusions, as I've said earlier, and the biggest one of all is the assumption that with Fermi NVIDIA is actually deemphasizing gaming at the same time they're expanding GPGPU. Really? A company who, even admitting so in their keynote, has made their success on gaming, is making strides to abandon that market? Could it be, just maybe, that it only seems that way because this launch was focused on Tesla and not on gaming? That we've simply only heard about one side of this architecture in a deliberate effort to not hurt existing sales or give their gaming competition a heads-up? How about this, why don't we instead come to such conclusions after they've talked about the gaming side of Fermi, hm?
Then there's the performance speculation. This part I actually enjoy, because it's almost detective work, digging through sparse details and trying to make some deductions, but the important thing to keep in mind at all times is that this is what it is: speculation. Assumptions. Conjecture. It's all about people's opinions. It's not hard to see why this is happening; I mean how long has it been that we've actually gotten new architectures? Seems like forever. Again, the market is stagnating, us enthusiasts are hungry for the old days of cutthroat competition and month-by-month developments. We're starved. This is all we have to feed on. NVIDIA has teased us, but that's it. We have nothing concrete, and the only people that can derive anything are application programmers interested in GPGPU stuff.
It can get frustrating though. Discussions go from civil to juvenile in a heartbeat, and I have my own guilty moments of getting sucked into it. I'm just like the rest of them, and yet I'm only just mature enough to see what's the matter with that. One thing's for sure: Fermi is a new architecture, it's more efficient and complex than anything NVIDIA or anyone has done before. It's the closest thing to a CPU a GPU has ever gotten, and what that translates to in the gaming world we can't know, until the day of the final product launch. There's too many contributing factors, and too many questions to draw any conclusions on performance. We've been thrown a bone, and all we're left with is an even greater hunger.
